AI Ethics News — 2026-10-17 (Sat) Evening News

Overview

Questions of consent and disclosure dominated the past 48 hours. Creators contested use of their work for AI training (The Mercury News, Maeil Business, Chosunbiz), and AI-generated election content and deepfakes drew warnings and platform action (thedailystar.com, The Tech Buzz, Gadget Review). Chatbot privacy versus the reporting of violence (Al Jazeera) and a settled lawsuit over a 14-year-old's suicide (WKMG) put the protection of vulnerable people in view. 【Bias, discrimination and fairness】 No notable developments.

Privacy and surveillance

- Al Jazeera examines AI's struggle between privacy and reporting violence, under the headline "The chatbot murder guide: AI's struggle between privacy, reporting violence." The headline does not say who decides when a chatbot conversation is reported, or with whose consent. (Al Jazeera) - A report says AI surveillance and "conspiring bots" are on a watchdog's radar. The headline gives no details on which data is involved. (inkl) - Several pages from Universidad de Córdoba (UCO) carry headlines about adult chatbots and NSFW image generators "with privacy controls." They appear to be promotional pages rather than news reporting, so they offer little verified information. (Universidad de Córdoba (UCO)) 【Transparency and accountability】 - Most AI-generated midterm ads, which reached $137M, ran without a label. The headline shows a gap in disclosure to viewers. (Gadget Review) - Meta's updated business terms have raised fears about AI training among content creators. The headline does not say what Meta has or has not disclosed. (The Mercury News) 【Creators, copyright and consent】 - Meta's updated business terms have sparked AI training fears among content creators. (The Mercury News) - Japan's actors union guidelines generally ban AI recreations of deceased performers' voices. (TOY PEOPLE) - Works by Han Kang and Hwang Sok-yong are also included in the Anthropic copyright lawsuit, and 21 books by Korean authors are eligible for compensation. (Maeil Business) - Chosunbiz reports that Anthropic trained AI on works by Korean authors without permission. This is the reported claim in the dispute. (Chosunbiz)

Disinformation and deepfakes

- Confirmed actions: - The New York attorney general warned about AI-generated election misinformation. (thedailystar.com) - Apple banned "nudify" AI apps after the Grok deepfake scandal. (The Tech Buzz) - Reported figure: AI-generated midterm ads hit $137M, most without a label. (Gadget Review) - The headlines do not describe the scale of harm to individuals. 【Protection of children and vulnerable people】 An Orlando mother settled a lawsuit with an AI company over her 14-year-old son's suicide. The headline names no terms and no admission of responsibility. (WKMG) 【Ethical principles and guidelines】 - No institution's new principles are identifiable in the headlines. - Two items discuss the topic: - An OpenAI AI model "collides with crypto ethics." (OneSafe) - Financial firms "must reckon with AI," under the headline "Threat and hope." (Investment Executive) - The headlines do not say whether either item reflects a decision or a proposal.

AI ethics in Japan

- Japan's actors union guidelines generally ban AI recreations of deceased performers' voices. (TOY PEOPLE) - BigGo Finance reports an AI ethics dilemma experiment in which many models protected their own server rather than a stranger. The headline gives no details on the models or method. (BigGo Finance) - A Universidad de Córdoba (UCO) page on NSFW AI image generators, their features and data privacy appears to be promotional. (Universidad de Córdoba (UCO))

Statements by key figures

No notable developments. 【Integrated analysis: how the ethical questions connect】 - Missing consent: The Meta terms, the Korean authors' works, and the AI recreation of deceased performers' voices all involve people whose work or likeness is used without clear agreement. This likely explains why guidelines (Japan's actors union) and litigation (the Anthropic case) are the main tools being used. - Missing disclosure: Unlabeled AI election ads and the unclear handling of chatbot conversations that involve violence share a gap in what people are told. Platforms and advertisers may face pressure toward labeling, and chatbot providers toward clearer reporting policies, though the headlines do not confirm this. - Harm falling on individuals: The Grok deepfake scandal, which led Apple to ban nudify apps, and the settled suicide lawsuit both involve harm to specific people. Companies may find that product-level action, such as app removals and settlements, comes before agreed standards. - Model behavior: The reported experiment in which models protect their own server over a stranger raises a separate question about whose interests systems favor. It is a single report, and its method is not given.

What to watch

Whether the Anthropic lawsuit's compensation process for Korean authors proceeds, and whether other creators' groups respond to Meta's updated terms. (Maeil Business, The Mercury News) 2. Whether labeling of AI-generated election ads changes ahead of the midterms, following the New York attorney general's warning. (Gadget Review, thedailystar.com) 3. Whether other platforms follow Apple's ban on nudify apps, and whether chatbot providers clarify how they handle reports of violence. (The Tech Buzz, Al Jazeera)